Ecological Economics 35 (2000) 47 – 61
                                               www.elsevier.com/locate/ecolecon




                          SPECIAL ISSUE

    THE VALUES OF WETLANDS: LANDSCAPE AND INSTITUTIONAL
             PERSPECTIVES
Valuing the environment as input: review of applications to
         mangrove-fishery linkages
                        Edward B. Barbier *
   Director, Centre for En6ironment and De6elopment Economics, En6ironment Department, Uni6ersity of York, Heslington,
                          York YO1 5DD, UK




Abstract

 The following paper reviews recent developments in the methodology for valuing the role of wetlands in supporting
economic activity. The main focus will be on mangroves serving as a breeding ground and nursery habitat in support
of coastal and marine fisheries. As this particular ecological function of a mangrove system means that it is effectively
an unpriced ‘environmental’ input into fisheries, then it is possible to value this contribution through applying the
production function approach. The first half of the paper overviews the procedure for valuing the environment as an
input, applied to the case of a wetland supporting a fishery. Both the ‘static’ Ellis – Fisher – Freeman approach and the
‘dynamic’ approach developed by Barbier and Strand, incorporating the intertemporal bioeconomic fishing problem,
are reviewed. The second half of the paper discusses briefly two recent case studies of mangrove-fishery valuation. An
application in South Thailand, which is based on the static Ellis – Fisher – Freeman model, and an application in
Campeche, Mexico, which is based on the dynamic approach. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Ecological functions; Environmental valuation; Fisheries; Habitat-fishery linkages; Mangroves


1. Introduction                           economic value of coastal wetland habitats in
                                  support of marine fisheries and other ecological
 The following paper overviews the general             functions, such as determining the value of marsh-
methodology for valuing mangrove-fishery link-            lands as habitat for Gulf Coast fisheries in the
ages that can be applied to a variety of mangrove          southern United States (Lynne et al., 1981; Ellis
and coastal wetland systems found around the            and Fisher, 1987; Farber and Costanza, 1987;
world. This approach has been used to assess the          Bell, 1989; Freeman, 1991; Bell, 1997) and the
                                  value of mangroves for coastal and marine
                                  fisheries in Thailand (Sathirathai, 1997) and Mex-
                                  ico (Barbier and Strand, 1998). This approach is
 * Tel.: +44-1904-434060; fax: +44-1904-432998.
                                  consistent with other related studies attempting to
 E-mail address: eb9@york.ac.uk (E.B. Barbier).

0921-8009/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 1 - 8 0 0 9 ( 0 0 ) 0 0 1 6 7 - 1
                E.B. Barbier / Ecological Economics 35 (2000) 47–61
48




                     Fig. 1. Valuing wetland benefits.


                              Here, the concern is mainly with indirect use
analyze habitat-fishery problems more generally,
                              values, i.e. the indirect support and protection
such as analyzing the competition between man-
                              provided to economic activity and property by a
groves and shrimp aquaculture in Ecuador (Parks
                              wetland’s natural ‘services’, or regulatory ecologi-
and Bonifaz, 1994), determining the value of a
                              cal functions. The ecological function of particu-
multiple-use mangrove system under different
                              lar interest is the role of a mangrove or coastal
management options in Bintuni Bay, Irian Jaya,
                              estuarine wetland system in serving as a breeding
Indonesia (Ruitenbeek, 1994), and examining gen-
                              ground or nursery for off-shore fisheries.
eral coastal system trade-offs, such as the effects
                                The main technique for valuing this ecological
of development and/or pollution on habitat-fish-
                              function of a wetland has been called, variously,
ery linkages (Kahn and Kemp, 1985; Knowler et
                              the production function approach, valuing the
al., 1997 Strand and Barbier, 1997; Strand and
                              environment as input and the value of changes in
Bockstael, 1990; Swallow, 1990; Swallow, 1994).
                              productivity approach (Freeman, 1991; Maler,  ¨
  Natural wetlands, including mangroves,
                              1991; Barbier, 1994). The basic assumption of this
provide many important functions for hu-
mankind, which can be grouped in terms of direct
use, indirect use and non-use values. Fig. 1 sum-
marizes the standard techniques available for as-       1
                                 For a guide to economic valuation of wetlands, see Barbier
sessing the various economic values of wetlands.1      et al. (1997).
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61                    49


approach is that, because the wetland serves as a      tems, such as coastal wetlands and mangroves,
breeding ground or nursery for the fishery, this       that protect or support economic activity, in par-
function can be treated as an additional environ-      ticular coastal and marine fisheries. This is the
                               production function approach.2
mental ‘input’ into the fishery. In static ap-
proaches, the welfare contribution of this input is       The general approach consists of a two-step
determined through producer and consumer sur-        procedure. First, the physical effects of changes in
plus measures of changes in the market equi-         a biological resource or ecological function on an
librium for harvested fish. In dynamic            economic activity are determined. Second, the
approaches, the wetland support function is in-       impact of these environmental changes is valued
cluded in the intertemporal bioeconomic harvest-       in terms of the corresponding change in the mar-
ing problem, usually as part of the growth          keted output of the corresponding activity. In
function of the fish stock, and any welfare im-        other words, the biological resource or ecological
pacts of a change in this function can be deter-       function is treated as an’input’ into the economic
mined in terms of changes in the long-run          activity, and like any other input, its value can be
equilibrium conditions of the fishery or in the        equated with its impact on the productivity of any
harvesting path to this equilibrium.             marketed output.
  The following section reviews both the static        More formally, if Q is the marketed output of
and dynamic production function approaches and        an economic activity, then Q can be considered to
their suggested applications to the wetland-fishery      be a function of a range of inputs:
valuation problem. Two recent case studies of
                               Q= F(Xi …Xk, S)                       (1)
valuing mangrove-fishery linkages are then re-
viewed. One applies the static methodology in          For example, the ecological function of particu-
Southern Thailand (Sathirathai, 1997), and the        lar interest is the role of mangroves in supporting
other applies the dynamic model in Campeche,         off-shore fisheries through serving both as a
Mexico (Barbier and Strand, 1998).              spawning ground and a nursery for fry. The area
                               of mangroves in a coastal region, S, may therefore
                               have a direct influence on the catch of mangrove-
2. The production function approach             dependent species, Q, which is independent from
                               the standard inputs of a commercial fishery,
  When a wetland is being indirectly used, in the
sense that the ecological functions of the wetland
are effectively supporting or protecting economic        2
                                 The production function approach discussed here is related
activity, then the value of these functions is essen-    to the household production function approach, which is a
                               more appropriate term for those surrogate market valuation
tially nonmarketed. However, economists have
                               techniques based on the derived demand by households for
demonstrated that it is possible to value such
                               environmental quality. That is, by explicitly incorporating
nonmarketed environmental services through the        non-marketed environmental functions in the modelling of
use of surrogate market valuation, which essen-       individuals’ preferences, household expenditures on private
tially uses information about a marketed good to       goods can be related to the derived demand for environmental
                               functions (Bockstael and McConnell, 1981; Freeman, 1993;
infer the value of a related nonmarketed good.
                               Smith, 1991). Some well-known techniques in applied environ-
Travel cost methods, recreational demand analy-
                               mental economics — such as travel cost, recreation demand,
sis, hedonic pricing and averting behaviour mod-       hedonic pricing and averting behaviour models — are based
els are all examples of surrogate market valuation      on the household production function approach. The dose-re-
that attempt to estimate the derived demand by        sponse technique is also related to the production function and
                               household production function approaches; however, dose-re-
households for environmental quality.
                               sponse models are generally used to relate environmental
  The following section describes another type of
                               damage (i.e. pollution, off-site impacts of soil erosion) to loss
surrogate market valuation that is particularly       of either consumer welfare (i.e. through health impacts) or
useful for the valuation of nonmarketed values        property and productivity (i.e. through damage to buildings,
associated with biological resources and ecosys-       impacts on production).
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61
50


Xi …Xk. Including mangrove area as a determinant       atic. In particular, assumptions concerning the
                               ecological relationships among these various multi-
of fish catch may therefore ‘capture’ some element
                               ple uses must be carefully constructed. Two major
of the economic contribution of this important
                               problems are double counting and trade offs be-
ecological support function.
                               tween various direct and indirect use values, which
  The above production function approach could
                               appear whenever analysts attempt to aggregate the
be applied potentially to any of the various indirect
                               various direct and indirect use values arising from
use values of wetland systems indicated in Fig. 1.
                               multiple use resource systems.
Thus this approach should prove to be a useful
                                Aylward and Barbier (1992) provide an example
method of estimating these nonmarketed — but
                               of both on-site and off-site double-counting in terms
often significant — economic values. However, in
                               of the nutrient retention function of a coastal
order for this method to be applied, it is extremely
                               wetland. Coastal wetlands often absorb organic
important that the relationship between any envi-
                               nutrients from sewage and other waste emitted into
ronmental regulatory function and the economic
                               waterways further upstream. Suppose that the
activity it protects or supports is well understood.
                               nutrients held by the wetland are indirectly support-
  Maler (1991) distinguishes between applications
   ¨
                               ing both shrimp production within the wetland area
of the production function approach. When produc-
                               and the growth of fish fry that supply an off-shore
tion, Q, is measurable and either there is a market
                               fishery. If the full value of the shrimp production
price for this output or one can be imputed, then
                               is already accounted for as a direct use value of the
determining the marginal value of the resource is
                               wetland’s resources, adding in the share of the
relatively straightforward. If Q cannot be measured
                               nutrient retention service as an indirect value and
directly, then either a marketed substitute has to be
                               aggregating these values would double count this
found, or possible complementarity or substitutabil-
                               indirect use. In other words, the value of shrimp
ity between S and one or more of the other
                               production already ‘captures’ the value-added con-
(marketed) inputs, Xi...Xk, has to be specified
                               tribution of nutrient retention.3 If instead one
explicitly. Although all these applications require
                               wanted to explicitly account for the value-added
detailed knowledge of the physical effects on pro-
                               contribution to shrimp production of the nutrient
duction of changes in the resource, S, and its
                               retention function, then the direct value of the
environmental functions, applications that assume
                               shrimp must be decreased to account for the return
complementarity or substitutability between the
                               in value now attached to the nutrient retention
resource and other inputs are particularly stringent
                               service.
on the information required on physical relation-        Similarly, if the fish fry supported through nutri-
ships in production. Clearly, cooperation is required    ents retained in the wetland eventually migrate to
between economists, ecologists and other re-         an off-shore fishery, then the indirect contribution
searchers to determine the precise nature of these      during the fry’s stay in the wetland is included as
relationships.                        on off-site component of the service’s value. That
  Applications of the production function ap-        is, the nutrient retention function of the wetland
proach may be most straightforward in the case of      produces an ‘external’ benefit in terms of supporting
single use systems, i.e. resource systems in which the    an off-shore fishery. Again, care must be taken to
predominant economic value is a single regulatory      adjust the value of harvested fish in any companion
function, or a group of ecological functions provid-     analysis of the adjoining fishery to avoid misrepre-
ing support or protection for an economic activity      senting the total economic value of the wetland and
in concert. In the case of multiple use systems —      the fishery taken together.
i.e. resource systems in which a regulatory function      Tradeoffs between two or more indirect use values
may support or protect many different economic        of a given ecosystem may also occur. For example,
activities, or which may have more than one
regulatory ecological function of important eco-        3
                                 On the other hand, if the nutrient retention function of the
nomic value — applications of the production         wetland is valued only by its contribution to shrimp produc-
function approach may be slightly more problem-       tion, this function would be undervalued.
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61                  51


Barbier et al. (1993) illustrate why it is necessary to    equilibrium. The second is essentially a dynamic
account for such trade-offs in their analysis of the     approach, which attempts to model the effects of
Hadejia–Jama’are floodplain in northern Nigeria.        a change in mangrove area on the growth function
The floodplain supports a number of important         of the intertemporal fishing problem. By solving for
agricultural, forestry and fishing activities within the    the long-run equilibrium of the fishery, the compar-
area of natural flooding. The floodplain also con-       ative static effects and resulting welfare impacts of
tributes to the recharge of groundwater, which is       a change in mangrove area on the equilibrium levels
in turn drawn off by numerous small village wells       of stock, effort and harvest can be determined.
throughout the region for domestic use and agricul-
tural activities. However, concerns have recently
been expressed about the excessive water use of        3. Static models
pump-irrigated wheat production within the flood-
plain. Increasing use of the floodplain water to         The static approach to valuing wetland-fishery
support this activity may mean less water available      linkages owes its development to a number of studies
for natural groundwater recharge, and thus for        that have tried to determine the value of marshlands
village wells outside the floodplain. If there are       as habitat for Gulf Coast fisheries in the southern
tradeoffs between the two environmental support        United States (Lynne et al., 1981; Ellis and Fisher,
functions, then adding the full value of the wetland’s    1987; Farber and Costanza, 1987; Bell, 1989; Free-
contribution to pump-irrigated wheat production        man, 1991; Bell, 1997).
within the floodplain to the full value of groundwa-       The initial method was first developed by Lynne
ter recharge of wells in neighbouring regions would      et al. (1981). Their approach was essentially half-
overestimate the total benefit of these two environ-      way between the ‘static’ and ‘dynamic’ approaches
mental functions. In fact, in their analysis the       described in this paper. Lynne et al. suggested that
authors had to adjust their estimates of the flood-      the support provided by the marshlands of southern
plain benefits for the ‘unsustainability’ of much       Florida for the Gulf Coast fisheries could be
pump-irrigated wheat production within the flood-       modelled by assuming that marshland area is an
ing area. The results of the analysis suggest that,      additional argument in the bioeconomic growth
even without considering the economic benefits of       equation of the fishery. Assuming that the latter
the groundwater recharge function, diverting water      function is logistic, and that harvesting of fish offsets
for upstream development does not make much          any natural growth in fish stock, then the authors
economic sense if it is detrimental to the natural      obtain the following relationship between fish har-
flooding system downstream.                  vest, h, fishing effort, E, and marshland area, M
  Despite these pitfalls, many recent studies have
                               ht = i0 + i1Et ln Mt − 1 + i2E 2 ln Mt − 1 + vt      (2)
                                               t
attempted to employ the production function ap-
proach in valuing one or more regulatory functions       The parameters of Eq. (2) can be estimated from
of wetlands, in particular the role of estuarine       data on harvest, fishing effort and marshland area
wetlands and mangroves in supporting off-shore        for those wetland-dependent species for which such
                               data are available.4 Lynne et al. use such estimates
fisheries. There are two ways in which this approach
has been implemented. The first is essentially a static    and the price of harvested fish to derive the value
approach, which either ignores the intertemporal
fish harvesting process (i.e. assumes single-period or      4
                                 Note that Eq. (2) does not represent the complete long-run
static production) or assumes that fish stocks are       equilibrium of a typical intertemporal fishing model as the
always constant (i.e. harvesting always offsets any      equation represents only one equilibrium condition, the bioe-
                               conomic condition of a constant level of fish stock. That is,
natural growth in the fish population). Either
                               because Eq. (2) excludes any consideration of price and costs
assumption can be used to derive a market equi-
                               in the determination of h, it does not represent the full
librium for fish harvest, and thus to estimate changes     long-run economic harvesting equilibrium of the fishery. For
in consumer and producer surplus arising from the       comparison, see the dynamic production function analysis
impacts of a change in mangrove area on this static      discussed in the next sub-section.
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61
52




        Fig. 2. Welfare measures in optimally managed and open access fisheries in static models.


of the marginal productivity of a change in wet-        one can employ this production function in a
land area in terms of h. For example, for the blue       standard static optimization model of profit-maxi-
crab fishery in western Florida salt marshes, the        mizing harvesting behaviour. This is essentially
authors obtain a marginal productivity of 2.3 lb        the approach adopted by Ellis and Fisher (1987),
per year for each acre of marshland. Others have        who use the Lynne et al. (1981) case study to
applied the Lynne et al. approach and Eq. (2) to        value the impacts of changes in the Florida Gulf
additional Gulf Coast fisheries in western Florida       Coast marshlands on the commercial blue crab
(Bell, 1989, 1997; Farber and Costanza, 1987).         fishery. Taking the sum of consumer and pro-
 However, it is possible to view Eq. (2) as a kind      ducer surplus as the measure of economic value,
of wetland-effort production function for a fish-        they hypothesize that an increase in wetland area
ery, and assuming a static or one-period model,        increases the abundance of crabs and thus lowers
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61             53


the cost of catch (see Fig. 2). The value of the       access and zero producer surplus, any reduction
wetlands’ support for the fishery, which in this       in the price of fish associated with the average
case is equivalent to the value of increments to       cost curve shifting down (in response to an in-
wetland area, can then be imputed from the re-        crease in wetland area) results in a gain in con-
sulting changes in consumer and producer           sumer surplus and increased wetland value.
surplus.                            Freeman also calculates the social value of the
 An important assumption in the Ellis and          marginal product of marshland area, VMPM,
Fisher model is that Lynne et al.’s Eq. (2) can be      which from Eq. (3) is
approximated by the Cobb – Douglas form
                                       h
                               VMPM = Pi                    (5)
h= AE aM b                     (3)            M
where h is the quantity of crab catch in pounds, E      where P is the price of crabs. As optimal regula-
is catch effort measured by traps set and M is        tion should lead to a higher price than open
area of wetlands. From the profit-maximizing         access, an inelastic demand means that VMPM is
conditions of the static optimization model for       higher under optimal regulation.
harvesting, the corresponding cost function, C, is       These different impacts of market conditions
                               and regulatory policies for the production func-
C = WA − 1/aM − i/hh 1/a              (4)
                               tion approach to valuing biological resources and
where W is the unit cost of effort. Assuming an       systems, where open access exploitation and im-
iso-elastic demand for crabs and either private       perfect markets for resources are common. As
ownership or optimal public management (i.e.         argued by Barbier (1994), this may be a prevalent
price equals marginal cost in both cases), Ellis and     feature of resource systems in tropical regions.
Fisher are able to estimate the change in con-         Applications of the production function ap-
sumer and producer surplus in the market for         proach to value more than one regulatory func-
blue crabs resulting from a change in marshland       tion of a wetland that supports or protects many
area (see Fig. 2).                      different economic activities are rare. As noted
  Freeman (1991) extends further Ellis and Fish-      above, assumptions concerning the ecological re-
er’s approach to show how the values imputed to       lationships among these various multiple uses
the wetlands are influenced by the market condi-       must be carefully constructed, and the data for
tions and regulatory policies that affect harvesting     this analysis are often not available.
decisions in the fishery, in particular whether it        For example, Ruitenbeek (1994) uses a
operates under conditions of open access or opti-      modified production function approach to evalu-
mal management. For example, under open ac-         ate the trade-offs between different forestry op-
cess, rents in the fishery would be dissipated, and      tions for a mangrove system in Bintuni Bay, Irian
price would be equated to average and not mar-        Jaya, Indonesia. The options range from preserv-
ginal costs. As a consequence, producer surplus is      ing the mangroves through a cutting ban to vari-
zero and only consumer surplus determines the        ous forestry development options involving
value of increased wetland area (see Fig. 2). Free-     partial, selective and clear cutting operations. An
man demonstrates that when the demand for          important feature of the analysis is that tries to
crabs is inelastic, the social value of an increase in    incorporate explicitly the linkages between loss of
area is higher under open access than under opti-      mangroves and their ecological functions and the
mal regulation, whereas the wetlands are more        productivity of economic activities. For example,
valuable under optimal regulation when demand        the mangroves may support many economic activ-
is elastic. This result stems from the role of price     ities, such as commercial shrimp fishing, commer-
changes in allocating welfare gains between pro-       cial sago production and traditional household
ducers and consumers: in the case of optimal         production from hunting, fishing, gathering and
regulations, part of the consumers’ gain is a trans-     cottage industry; they may also have an indirect
fer from producer surplus, whereas under open        use value through controlling erosion and sedi-
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61
54


                               4. Dynamic models
mentation, which protects agricultural production
in the region; and they have an indirect role in
                                The production function approach can also be
supporting biodiversity. To the extent that the
                               incorporated into intertemporal models of renew-
ecological linkages in terms of support or protec-
                               able resource harvesting in cases where the eco-
tion of these activities are strong, then the oppor-
                               logical function affects the growth rate of a stock
tunity cost of forestry options that lead to the
                               over time. In such cases, the production function
depletion or degradation of the mangroves will be
                               link is a dynamic one, as the ecological function
high. Thus, the ‘optimal’ forest management op-
                               affects the rate at which a renewable resource
tion — whether clear cutting, selective cutting or
                               increases over time, which in turn affects the
complete preservation — depends critically on the
                               amount of offtake, or harvest, of the resource.
strength of the ecological linkages.
                               The basic approach to valuation of an environ-
  In the absence of any ecological data on these
                               mental input to renewable resource production in
linkages, Ruitenbeek develops several different
                               a dynamic context is outlined by Hammack and
scenarios based on different linkage assumptions.
                               Brown (1974), Ellis and Fisher (1987), Freeman
This essentially amounted to specifying more spe-
                               (1993), Barbier and Strand (1998).
cifically the relationship between Q and S in the
                                As shown by Barbier and Strand (1998), adapt-
simple production function relationship Eq. (1)
                               ing bioeconomic fishery models to account for the
indicated above. Thus for each productive activity
                               role of a mangrove system in terms of supporting
at time t, Qit, the following relationship is
                               the fishery as a breeding ground and nursery
assumed
                               habitat is fairly straightforward, if it is assumed in
Qit /Qi0 =(St − ~ /S0)h              (6)    the fishery model that the effect of changes in
                               mangrove area is on the carrying capacity of the
where St is the area of remaining undisturbed
                               stock and thus indirectly on production.5 Defining
mangroves at time t, h and ~ are impact intensity
                               Xt as the stock of fish measured in biomass units,
and delay parameters, respectively, Qi0 =Qit (t =
                               any net change in growth of this stock over time
0) and S0 =St (t= 0). For example, for fishery-
                               can be represented as
mangrove linkages, a moderate linkage of h =0.5
                               Xt + 1 − Xt = F(Xt, Mt )− h(Xt, Et ), FX \ 0, FM
and ~=5 would imply that shrimp output varies
with the square root of mangrove area (e.g. a 50%             \0                     (7)
reduction in mangrove area would result in a 30%
                                Thus net expansion in the fish stock occurs as a
fall in shrimp production), and there would be a
                               result of biological growth in the current period,
delay of 5 years before the impact takes effect. If
                               F(Xt, Mt ), net of any harvesting, h(Xt, Et ). Note
no ecological linkages are present, i.e. there is no
                               that the standard fish harvesting function is em-
indirect use value of mangroves in terms of sup-
                               ployed; i.e. harvesting is a function of the stock as
porting shrimp fishing, then h = 0. At the other
                               well as fishing effort, Et. Instead, it is the biologi-
extreme, very strong linkages imply that the im-
                               cal growth function of the fishery that is modified
pacts of mangrove removal are linear and imme-
                               to allow for the influence of mangrove area, Mt,
diate, i.e. h =1 and ~ =0.
                               as a breeding ground and nursery. It is reasonable
  Based on his analysis, Ruitenbeek concludes
                               to assume that this influence on growth is posi-
that the assumption of no or weak environmental
                               tive, i.e. (F/(Mt = FM \ 0, as an increase in man-
linkages is unrealistic for most economic activities
                               grove area will mean more carrying capacity for
related to the mangroves. Moreover, given the
                               the fishery and thus greater biological growth.
uncertainty over these ecological linkages and the
high costs associated with irreversible loss, if envi-
ronmental linkages prove to be significant, then
only modest selective cutting (e.g. 25% or less) of      5
                                 For analytical convenience, a discrete time model of the
the mangrove area was recommended.              fishery is employed here.
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61              55




           Fig. 3. Mangrove loss and the long-run equilibrium of an open access fishery.

  Eq. (7) can now be employed in a standard         shows an optimal path to a stable long-run equi-
intertemporal harvesting model of the fishery,         librium for the fishery. In this case, a decrease in
where depending on the management regime, har-        mangrove area causes the long-run level of fishing
vesting over time can either be depicted to occur       effort to fall. As harvesting levels are generally
under open access conditions (i.e. effort in the       positively related to effort levels, the consequence
fishery adjusts over time to the availability of        of mangrove deforestation is also a decrease in
profits) or under optimal management conditions        equilibrium fish harvest.
(the discounted net returns from harvesting the
fishery are maximized over time). The effect of a
change in mangrove area can therefore be valued        5. A case study of a static model: southern
in terms of changes in the optimal path of harvest-      Thailand
ing over the period of analysis and in terms of the
changes in the long-run equilibrium of the fishery.        Sathirathai (1997) uses the Ellis-Fisher-Freeman
  Fig. 3 shows the fairly straightforward case        model to value the welfare impacts of mangrove
analyzed by Barbier and Strand, where the effects       deforestation on coastal fisheries in Surat Thani
of a change in mangrove area is depicted in terms       Province on the Gulf of Thailand. In recent
of influencing the long-run equilibrium of an open       decades, the expansion of intensive shrimp farm-
access fishery. In the figure, the long-run equi-        ing in the coastal areas of southern Thailand has
librium of the fishery is depicted in terms of steady     led to rapid conversion of mangroves. Over 1975–
values for effort, E, and fish stocks, X. As dis-       1993 the area of mangroves has virtually halved,
cussed above, the carrying capacity of the fishery       from 312 700 hectares (ha) to 168 683 ha. Al-
is assumed to be an increasing function of man-        though the rate of mangrove loss has slowed, in
grove area, i.e. K=K(M), KM \0. Trajectory one        the early 1990s the annual loss was estimated to
                    E.B. Barbier / Ecological Economics 35 (2000) 47–61
56

Table 1
Welfare estimates of changes in mangrove area on the Gulf of Thailand shellfish and demersal fisheriesa

                  Economic value of a change in mangrove area (US$ per ha)b

Management regime         Demand elasticity        Demersal fish        Shellfish       All fish

(Open access)
                  p= −0.1               63.48           46.75         110.23
                  p= −1                39.71           43.29         83.00
                  p= −10                8.38           24.92         33.30
(Managed fisheries)
                  p= −0.1               38.74           44.47         83.21
                  p= −1                38.88           44.50         83.38
                  p= −10               39.06           44.63         83.69
                  Economic value of annual loss of 1200 ha of mangrove area (US$)c

(Open access)
                  p= −0.1             76 176           56 100        132 276
                  p= −1              47 652           51 948         99 600
                  p= −10             10 056           29 904         39 960
(Managed fisheries)
                  p= −0.1             46 488           53 364         99 852
                  p= −1              46 656           53 400        100 056
                  p= −10             46 872           53 556        100 428

 a
  Based on Sathirathai (1997).
 b
  Calculations assume an initial equilibrium quantity demand and price based on observed data for Surat Thani Province (Zone
three) in 1993. For demersal fish this is harvested output of 1 545 000 kg and price of US$ 1.51/kg, and for shellfish 1 917 000 kg
and US$2.58/kg.
 c
  Over 1991–3, the average annual loss of mangroves in Surat Thani province (Zone three) was estimated to be around 12.19 km2,
or around 1200 hectares (ha).

be around 3000 ha/year for all of Thailand, and           A separate harvesting function is assumed to ap-
                                  ply to demersal fish as opposed to shellfish.
1200 ha/year in Surat Thani province.
                                    Sathirathai conducts a panel analysis to esti-
  The Gulf of Thailand mangroves are thought to
                                  mate a log-linear version of Eq. (3) for all shellfish
provide breeding grounds and nurseries in sup-
                                  and all demersal fish in the Gulf of Thailand. The
port of several species of demersal fish and
shellfish, mainly crab and shrimp.6 To analyze the          analysis combines harvesting, effort and man-
                                  grove data across all five zones of the Gulf of
impact of mangrove deforestation on these
                                  Thailand and over the 1983–93 time period. This
fisheries in Surat Thani, Sathirathai assumes that
                                  allows estimation of the parameters A, h and i in
harvesting in both demersal and shellfish fisheries
                                  Eq. (3), for two separate Cobb–Douglas produc-
is a Cobb–Douglas function of the level of fishing
effort and mangrove area, as depicted by Eq. (3).7         tion functions, one each for demersal fish and
                                  shellfish. Combining this information with the
                                  estimated unit cost of effort, W, allows the
 6
  Mangrove-dependent demersal fish include those belonging
                                  Cobb–Douglas cost function Eq. (4) to be spe-
to the Clupeidae, Chanidae, Ariidae, Pltosidae, Mugilidae, Lu-
janidae and Latidae families. The shellfish include those be-    cified for both demersal fish and shellfish for each
longing to the families of Panaeidae for shrimp and Grapsidae,   of these fisheries in Surat Thani Province. This
Ocypodidae and Portnidae for crab.
                                  province is an important fishing region in Zone
 7
  In this study, total fishing effort per year is the number of
                                  three of the Gulf of Thailand. Following the
fishing instruments (e.g. gill net boats) recorded per anum
                                  methodology indicated in Fig. 2, Sathirathai uses
times the average of hours spent on fishing per fishing instru-
                                  the cost functions derived for each fishery to
ment each year.
                E.B. Barbier / Ecological Economics 35 (2000) 47–61              57


estimate the likely welfare impacts of a change in     the role of mangroves in the Laguna de Terminos
mangrove area in Surat Thani, assuming alterna-       in supporting the shrimp fishery of Campeche,
tively open access and managed fishery conditions.      Mexico.
  Table 1 shows the results of the welfare calcula-      Mexico’s gulf coast states account for over half
tion for the impact of a per ha change in mangrove     of the country’s shrimp catch, and the state of
area on the Gulf of Thailand shellfish and demersal     Campeche is responsible for one-sixth of Mexico’s
fisheries of Surat Thani Province. For all man-       total output of shrimp. Campeche’s shrimp fishery
grove-dependent fisheries, the value of a change in     employs about 13% of the state’s economically
mangrove area ranges from US$33-110/ha, de-         active population. In recent years the total number
pending on whether the fisheries are open access or     of boats in the fishery have increased substantially,
managed. Similar to the outcome reported by         but the composition of the fleet has also changed
Freeman (1991) for the Florida Gulf Coast blue       significantly. There has been a substantial decline
crab fishery, when the demand for Gulf of Thailand      in the number of commercial vessels, whereas the
fish is inelastic, the value of a change in mangrove     artisanal fleet has expanded rapidly. From 1980–
area in Surat Thani is higher under open access       1987, production in the shrimp fishery fluctuated
than under optimal regulation, whereas the wet-       steadily between 7–8 thousand metric tons (KMT),
lands are more valuable under optimal regulation      but by 1990 output had fallen to 4.6 KMT.
when demand is elastic. Under managed fishery          The mangroves in the Laguna de Terminos are
conditions, different demand elasticity assumptions     considered by ecologists to be the main breeding
hardly affect the welfare estimates of a change in     ground and nursery habitat for the shrimp fry of
mangrove area, which are estimated to be around       the Campeche fishery (Yanez-Arancibia and Day,
                                            ˜
US$39/ha for demersal fish and US$45/ha for         1988). Mangrove area was estimated to be around
                              860 km2 in 1980, declining to about 835 km2 in
shellfish. In the open access scenario, changes in
                              1991, a loss of around 2 km2 per annum. The
elasticities affect more the value of mangroves in
supporting demersal fisheries as opposed to         primary reason for the loss is the encroachment of
shellfish. Thus under open access and depending on      population from Carmen, the large city adjacent to
the elasticity of demand, the value of the mangroves    Laguna de Terminos. Future threats are expected
in Surat Thani ranges from US$8-63/ha for demer-      to come from expansion of shrimp aquaculture
sal fish and from US$25-47/ha for shellfish.         through conversion of coastal mangroves, and
  Table 1 also shows the estimated welfare impacts     possibly pollution.
associated with the annual loss of 1200 ha of         Barbier and Strand model the effects of man-
mangroves in Surat Thani, which was approxi-        grove deforestation in Laguna de Terminos by use
mately the annual rate of mangrove conversion        of comparative static analysis of the long-run
recorded in the early 1990s in the province. Given     equilibrium, as depicted in Fig. 3. In their model
this rate of deforestation, the economic loss in      of the Campeche shrimp fishery, they assume that
terms of support of the Gulf of Thailand fisheries      the basic growth function of the fishery is logistic
in Zone three is estimated to be around           and that shrimp harvesting follows the conven-
US$100 000 per year, if these fisheries were opti-      tional Schaefer production process, ht = qEtXt.
mally managed. Under open access conditions, this      Thus Eq. (7) becomes
economic loss ranges from US$40 000 to 132 000,
                              Xt + 1 − Xt = [r(K(Mt )− Xt )− qEt ]Xt      (8)
depending on demand elasticities.
                              where r is the intrinsic growth of shrimp each
                              period, K is the environmental carrying capacity of
6. A case study of a dynamic model: Campeche,        the system and mangrove area, Mt, has a positive
                              impact on carrying capacity, i.e. KM \ 0.
Mexico
                               To estimate the comparative static effects of a
 Barbier and Strand (1998) employ the dynamic       change in mangrove area on long-run shrimp
approach to production function analysis to value      harvesting, Barbier and Strand assume a propor-
                    E.B. Barbier / Ecological Economics 35 (2000) 47–61
58


tional relationship between mangrove area and            (9) with the price and cost data, the authors
carrying capacity, i.e. K(M) = aM, h \ 0. As the           simulate the effects of a change of mangrove area
shrimp stock is constant in the long-run equi-            on equilibrium harvesting and gross revenues in
librium, Xt = Xt + 1 = X, then using this condition         the Campeche shrimp fishery over 1980–90.
in Eq. (8) and the Schaefer production function to           The results are depicted in Table 2. On average
                                   over the 1980–90 period, a marginal (in km2)
substitute for X, the following relationship be-
tween shrimp production, mangrove area and ef-            decline in mangrove area produces a loss of about
fort is derived                           14.4 metric tons of shrimp harvest and nearly
                                   US$140 000 in revenues from the Campeche fish-
          q2 2    q2
           E =qhEM − E 2
h=qEK(M)−                         (9)    ery each year. However, given the relatively small
          r     r
                                   rate of annual mangrove deforestation in the re-
                                   gion over the 1980–90 period — around 2 km2
  The authors estimate Eq. (9) by employing
1980 –90 time series data on shrimp harvests,            per year — the resulting loss in shrimp harvest
effort and mangrove area for Campeche, Mexico,            and revenues does not appear to be substantial,
to derive the parameters b1 =hq and b2 = − q 2/r.          only around 0.4% per year.
  A second condition of the long-run equilibrium            The simulation in Table 2 also demonstrates
of an open access fishery is that profits will be           how the economic losses associated with man-
zero, i.e. ph =cE, where p is the price of shrimp          grove deforestation are affected by long-run man-
catch and c is the cost of fishing effort. In order to        agement of the open access fishery. As noted
simulate the comparative static effects of a change         above, the early years of the period of analysis
in mangrove area, Barbier and Strand assume             (e.g. 1980–81) were characterized by much lower
that this ‘zero profit’ condition holds for the            levels of fishing effort and higher harvests (e.g. on
Campeche shrimp fishery. Using actual price data           average around 4800 combined vessels extracting
on shrimp catches over this period, they calculate          about 8.5 KMT annually). Table 2 shows that, if
the costs of effort, c A, necessary for the zero profit        this earlier period represented the open access
condition to hold for the Campeche fishery over            equilibrium of the fishery, the economic impacts
                                   of a marginal (km2) decline in mangrove area
1980 –90. Using the estimated parameters of Eq.

Table 2
Simulation estimates of a marginal change in mangrove area, Campeche, Mexico (d M)a,b

              Cost (c A)
Year    Price (p)             Change in equilibrium harvest    Change in equilibrium     Change %
      US$/kgc     US$/vesseld    (d h A) metric tons         revenues (pd h A) US$

1980     7.10      13 984       20.40                144 808            0.23
1981     9.68      15 628       16.72                161 826            0.20
1982    10.57      13 816       13.53                143 060            0.18
1983     9.80      13 636       14.41                141 197            0.18
1984     9.83      14 096       14.85                145 963            0.19
1985     9.80      16 687       17.63                172 798            0.20
1986    10.00      15 013       15.55                155 460            0.19
1987    10.22      14 363       14.55                148 731            0.20
1988    10.56      14 132       13.86                146 334            0.20
1989    10.21      10 000       10.14                103 547            0.17
1990    10.40       6677       6.65                 69 143            0.14
Mean     9.83      13 457       14.39                139 352            0.19

 a
   Source: Barbier and Strand (1998).
 b
   Parameter estimates: b1 = 4.4491; b2 = −0.4297.
  c
   US$/kg, in real (1982) prices.
  dA
   c is the ‘equilibrium’ (real) cost per unit effort, defined as the cost level necessary to attain zero profit in the fishery, i.e.
c A =ph A/E A.
                 E.B. Barbier / Ecological Economics 35 (2000) 47–61              59


would be a reduction in annual shrimp harvests of      namic approaches to valuing the support function
around 18.6 tons, or a loss of about US$153 300       of mangroves in Thailand and Mexico have also
per year. In contrast, the last two years of the       been examined.
analysis (e.g. 1989–90) saw much higher levels of        The production function approach appears to
effort and lower harvests in the fishery (e.g.        be well suited to valuing the important ecological
around 6700 combined vessels extracting 5.3 KMT       role of coastal and estuarine wetlands in support-
annually). As a consequence, if this latter period      ing offshore fisheries. As these wetland systems are
represents the open access equilibrium, then a        under considerable threat from coastal develop-
marginal decline in mangrove area would result in      ment, it is important to develop reliable economic
annual losses in shrimp harvests of 8.4 tons, or       estimates of the value of their ecological support
US$86 345 each year.                     function role. Failure to consider this value may
  Thus, the value of the Laguna de Terminos         misrepresent the economic costs associated with
mangrove habitat in supporting the Campeche         wetland conversion, which are too often assumed
shrimp fishery appears to be affected by the level      to be insignificant or zero in coastal development
of exploitation. This suggests that, if an open       decisions. This is particularly the case in develop-
access fishery is more heavily exploited in the long     ing countries, where many mangrove systems are
run, the subsequent welfare losses associated with      threatened with conversion through the expansion
the destruction of natural habitat supporting this      in coastal areas of aquaculture, agriculture,
fishery are likely to be lower. Intuitively, this       tourism, and urban and infrastructural develop-
makes sense. The economic value of an over-ex-        ment.
ploited fishery will be lower than if it were less        However, both static and dynamic models show
heavily depleted in the long run. The share of this     that, in applying the production function ap-
value that is attributable to the ecological support     proach to valuing the support of wetlands for
function of natural habitat will therefore also be      offshore fisheries, any resulting welfare estimate
smaller.                           will be affected significantly by whether the
  The management implications are clear: As long      fisheries are managed or subject to open access.
as effort levels continue to rise, harvests will fall,    For example, the Gulf of Thailand study indicates
even if mangrove areas are fully protected. More-      that static production function estimates of the
over, any increase in harvest and revenues from an      value of a change in mangrove area in terms of
expansion in mangrove area is likely to be short-      support of managed fisheries will be little affected
lived, as it would simply draw more effort into the     by different demand elasticities. In contrast, for
fishery. Better management of the Campeche          open access fisheries, the value of the mangrove
shrimp fishery to control over-exploitation may be      support function will tend to be much lower for
the only short-term policy to bring production        elastic as opposed to inelastic market demand for
back to respectable levels, as well as realizing the     harvested fish. In the case of dynamic models of
more long-term economic benefits of protecting        mangrove-fishery linkages, the Mexican case study
mangrove habitat.                      illustrates how the economic losses associated with
                               mangrove deforestation are influenced by the
                               long-run management conditions in the open ac-
7. Conclusion                        cess fishery. That is, if an open access fishery is
                               more heavily exploited in the long run, the subse-
  This paper has indicated how the economic         quent welfare losses associated with any mangrove
value of mangroves in supporting coastal and         habitat supporting this fishery are likely to be
marine fisheries can be estimated through applica-      lower, as the mangroves will now be supporting a
tion of production function approaches. Both ba-       more over-exploited and thus less valuable fishery.
sic static and dynamic production function models        The methodologies and case studies discussed in
for estimating this value have been reviewed. Case      this paper show the important potential in utiliz-
studies of the application of the static and dy-       ing production function approaches to valuing the
                    E.B. Barbier / Ecological Economics 35 (2000) 47–61
60

                                  Bockstael, N.E., McConnell, K.E., 1981. Theory and estima-
environment as input, particularly valuing the
                                    tion of the household production function for wildlife
ecological support functions of wetlands, such as
                                    recreation. J. Environ. Econ. Manag. 8, 199 – 214.
mangrove systems. Ecologists have indicated that          Bell, F.W., 1989. Application of Wetland Valuation Theory to
the regulatory functions performed by wetlands             Florida Fisheries. Report No. 95, Florida Sea Grant Pro-
and other complex natural ecosystems may be               gram, Florida State University, Tallahassee.
highly significant in supporting and protecting           Bell, F.W., 1997. The economic value of saltwater marsh
                                    supporting marine recreational fishing in the Southeastern
economic activity. Perhaps the next phase in the
                                    United States. Ecol. Econ. 21, 243 – 254.
development of production function approaches           Ellis, G.M., Fisher, A.C., 1987. Valuing the environment as
will be to apply such methodologies not just to             input. J. Environ. Manag. 25, 149 – 156.
valuing single-use functions of wetlands, such as         Farber, S., Costanza, R., 1987. The economic value of wet-
the role of mangroves as nursery and breeding              lands systems. J. Environ. Manag. 24, 41 – 51.
                                  Freeman, A.M., 1991. Valuing environmental resources under
ground habitats for coastal and marine fisheries,
                                    alternative management regimes. Ecol. Econ. 3, 247 – 256.
but also to valuing simultaneously the diverse
                                  Freeman, A.M., 1993. The Measurement of Environmental
range of regulatory functions typically found in a           and Resource Values: Theory and Methods. Resources for
multi-use natural wetland, such as those listed in           the Future, Washington DC, pp. 516.
Fig. 1.                              Hammack, J. Brown, G.M. Jr., 1974. Waterfowl and Wet-
                                    lands: Towards Bioeconomic Analysis. Resources for the
                                    Future, Washington DC, pp 273.
                                  Kahn, J.R., Kemp, W.M., 1985. Economic losses associated
Acknowledgements
                                    with the degradation of an ecosystem: the case of sub-
                                    merged aquatic vegetation in Chesapeake Bay. J. Environ.
  A version of this paper was prepared for the             Econ. Manag. 12, 246 – 263.
4th Workshop of the Global Wetlands Economics           Knowler, D., Strand, I. Barbier, E.B., 1997. An Economic
                                    Analysis of Black Sea Fisheries and Environmental Man-
Network (GWEN), Wetlands: Landscape and In-
                                    agement. Final Report, The Black Sea Environment Pro-
stitutional Perspectives, Beijer International Insti-
                                    gramme, The World Bank/UNEP Global Enviromental
tute of Ecological Economics, The Royal Swedish             Facility, Istanbul, Rome.
Academy of Sciences, Stockholm, Sweden, 16 – 17          Lynne, G.D., Conroy, P., Prochaska, F.J., 1981. Economic
November 1997. I am grateful to Jack Ruiten-              value of marsh areas for marine production processes. J.
                                    Environ. Econ. Manag. 8, 175 – 186.
beek, Tore Soderquist and an anonymous referee
       ¨
                                  Maler, K.-G., 1991. The production function approach. In:
                                   ¨
for helpful comments. However, the usual caveat
                                    Vincent, J.R. Crawford, E.W. and Hoehn, J.P. (Eds.)
applies.                                Valuing Environmental Benefits in Developing Countries.
                                    Special Report 29. Michigan State University, East Lans-
                                    ing, pp. 11 – 32.
References                             Parks, P., Bonifaz, M., 1994. Nonsustainable use of renewable
                                    resources: mangrove deforestation and mariculture in
                                    Ecuador. Mar. Resour. Econ. 9, 1 – 18.
Aylward, B.A., Barbier, E.B., 1992. Valuing environmental
                                  Ruitenbeek, H.J., 1994. Modelling economy-ecology linkages
  functions in developing countries. Biodivers. Conserv. 1,
                                    in mangroves: economic evidence for promoting conserva-
  34– 50.
                                    tion in Bintuni Bay, Indonesia. Ecol. Econ. 10, 233 – 247.
Barbier, E.B., 1994. Valuing environmental functions: tropical
                                  Sathirathai, S., 1997. Economic Valuation of Mangroves and
  wetlands. Land Econom. 70, 155–173.
                                    the Roles of Local Communities in the Conservation of the
Barbier, E.B., Acreman, M. Knowler, D., 1997. Economic
                                    Resources: Case Study of Surat Thani, South of Thailand.
  Valuation of Wetlands: A Guide for Policymakers. Ramsar
                                    Final report submitted to the Economy and Environment
  Convention Bureau, Geneva, 127 pp.
                                    Program for Southeast Asia (EEPSEA), EEPSEA,
Barbier, E.B., Adams, W.M., Kimmage, K., 1993. An eco-
                                    Singapore.
  nomic valuation of wetland benefits. In: Hollis, G.E.,
                                  Smith, V.K., 1991. Household production functions and envi-
  Adams, W.M., Aminu-Kano, M. (Eds.), The Hadejia–
                                    ronmental benefit estimation. In: Braden, J.B., Kolstad,
  Nguru Wetlands: Environment, Economy and Sustainable
                                    C.D. (Eds.), Measuring the Demand for Environmental
  Development of a Sahelian Floodplain Wetland. IUCN,
                                    Quality. North-Holland, Amsterdam, pp. 41 – 76.
  Geneva, pp. 191 – 209.
                                  Strand, I.E., Bockstael, N.E., 1990. Interaction between agri-
Barbier, E.B., Strand, I., 1998. Valuing mangrove-fishery link-
                                    culture and fisheries: empirical evidence and policy implica-
  ages: a case study of Campeche, Mexico. Environ. Resour.
  Econ. 12, 151 – 166.                         tions. In: Just, R.E., Bockstael, N.E. (Eds.), Commodity
                   E.B. Barbier / Ecological Economics 35 (2000) 47–61                   61

  and Resource Policies in Agricultural Systems. Springer-     Swallow, S.K., 1994. Renewable and nonrenewable resource
  Verlage, New York, pp. 50–61.                    theory applied to coastal agriculture, forest, wetland and
Swallow, S.K., 1990. Depletion of the environmental basis for     fishery linkages. Mar. Resour. Econ. 9, 291 – 310.
  renewable resources: the economics of interdependent re-     Yanez-Arancibia, A., Day, J.W. Jr. (Eds.), 1988. Ecology of
                                   ˜
  newable and nonrenewable resources. J. Environ. Econ.        coastal ecosystems in the southern gulf of Mexico: The
  Manag. 19, 281 – 296.                        Terminos Lagoon Region. UNAM Press, Mexico, 518 pp.




                                .